Opc Errors

PackageReadError

_class

Thrown when the input is a ZIP container but it cannot be unpacked: the archive is corrupt or truncated, or inflating it would push total uncompressed output past the caller's bound (which is how a zip bomb presents).

The neighbouring UnsupportedFormatError says the input is a different kind of thing; this one says it is the right kind and we will not (or cannot) unpack it. Keeping them apart is what lets a caller answer "should I try another reader, or reject this file?", and it is what replaced the message-prefix match the bomb refusal used to be recognised by.

The zip library's own failure text never survives into either the message or cause: it can name internals, or an absolute filesystem path, from the layer below, and this type carries the classification precisely so no lower-layer string has to.

class PackageReadError extends XlsxError {
  override readonly name = 'PackageReadError';
  override readonly code = 'malformed-input';
}

---

UnsupportedFormat

_type

Which unsupported input the reader recognised:

• 'xls': a legacy BIFF .xls (an OLE2/CFB compound file), detected by its magic bytes.
• 'xlsb': a binary BIFF12 .xlsb, the same OPC/ZIP container as .xlsx, but its office document is xl/workbook.bin rather than xl/workbook.xml. readXlsx/readXlsb read one; the entry points that cannot yet (the row streamer) report it under this format with their own message.
• 'unknown': not a recognised spreadsheet at all, meaning not a ZIP, or a ZIP carrying no OOXML workbook part (nor a .xlsb binary one).

type UnsupportedFormat = 'xls' | 'xlsb' | 'unknown';

---

UnsupportedFormatError

_class

Thrown when input is not a readable .xlsx package. The single format field is the branch a caller keys on (rather than a subclass per format), so a catch can distinguish a legacy .xls, a binary .xlsb, and an unrecognised blob without string-matching the message.

The message never carries a filesystem path or the underlying zip library's internals. The whole point of the type is that the classification, not a leaked lower-layer string, is what the caller sees.

format stays the branch for which unsupported input this was; the inherited XlsxError.code answers the coarser question of what kind of failure it is.

class UnsupportedFormatError extends XlsxError {
  override readonly name = 'UnsupportedFormatError';
  override readonly code = 'unsupported-format';
  readonly format: UnsupportedFormat;
}